Angiographic assessment of coronary atherosclerosis is a definitive procedure for patient management and an endpoint in clinical trials. Computer methods promise improvements in accuracy, precision and characterization of estimates of lumen geometry. Since angiography and computer processing involve cost, it is important to quantitate these improvements not only to evaluate the benefit but also to provide a basis for sample size estimates of clinical trials. We will develop and test a series of quantitative computer atherosclerosis measures including relative stenosis, absolute stenosis, and cross-sectional area. The evaluation will be done using phantoms alone and phantom inserts into dogs. Operational feasibility will be evaluated by means of a user station with automation of vessel segment selection.